50 SOCIAL SCIENTIST of individual ranking, is somewhat devastating for economic theory. As corollaries, many theorems based on this implicit congruence appear to be in a hopeless confusion of categories—now talking about choice and in the next step deriving theorems on individual welfare. The most remarkable example would be the two classic optimality theorems of welfare economics that celebrate the correspondence between market equilibria and Pareto-optimality—where the equilibrium achieved on the basis of individual choice is branded as optimal in terms of welfare on the implicit use of this misplaced congruence. The author's arguments are however more constructive than just aimed at undermining the received theory. Once it is recognised that all agents can be better off in terms of their individual welfare ranking (as in a Prisoners5 Dilemma) by acting according to some "as if'5 ranking, there arises the possibility of analysing a whole range of behaviour which is traditionally left out by economists as beyond the scope of their discipline. Such diverse phenomena as behaviour based on commitment to a group, like the family, clan, caste, club or class or based on moral incentives for example, need not entirely belong to the mysterious territory of the sociologist or the political scientist. From here. Prof Sen departs in two related but distinct directions. One of them is to demand more non-utility information in the analysis of economic behaviour. Clearly in a wide class of behaviour the usual analysis of welfare maximisation fails for the simple reason that behaviour indeed has many other dimensions like motivation, commitment, morality, to name a few. Analytically however this need not create any problems if we could capture these elements in a single subjective ranking of the individual.2 The problem is that neither can the ranking obtained from behaviour through observed choices be given such an interpretation, nor can we even theoretically conceive of a single subjective ranking that could capture these additional elements in choice behaviour. Thus an argument develops against what Prof Sen calls "welfarism", the practice of explaining behaviour, individual and social, by reference to individual welfare alone, while the latter is just one element of the set of determinants of behaviour. The critique of "welfarism55 in the form in which it has come is important. The kind of non-utility information that will be necessary for analysis is neither a characteristic of the social states alone (in the way that utility or welfare itself is), nor of the agent^ subjective makeup. The categories like commitment, motivation or morality require for their definitional basis the entire information about the context of the choice problem. Thus the demand for more non-utility information is in effect a demand for posing the problem of choice not as an inward looking formulation of mechanical homocentrism, but to incorporate as much of the information about interaction between the agent and his social environment as possible within the traditional format of choice analysis.